首页|Study Results from Tsinghua University Broaden Understanding of Machine Learning (A Combined Active Control Method of Restricted Nonlinear Model and Machine Lea rning Technology for Drag Reduction In Turbulent Channel Flow)
Study Results from Tsinghua University Broaden Understanding of Machine Learning (A Combined Active Control Method of Restricted Nonlinear Model and Machine Lea rning Technology for Drag Reduction In Turbulent Channel Flow)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
A new study on Machine Learning is now available. According to news reporting originating in Beijing, People's Republi c of China, by NewsRx journalists, research stated, "The practical implementatio n of machine learning in flow control is limited due to its significant training expenses. In the present study the convolutional neural network (CNN) trained w ith the data of the restricted nonlinear (RNL) model is used to predict the norm al velocity on a detection plane at y(+) = 10 in a turbulent channel flow, and t he predicted velocity is used as wall blowing and suction for drag reduction." Financial support for this research came from National Natural Science Foundatio n of China (NSFC). The news reporters obtained a quote from the research from Tsinghua University, "An active control test is carried out by using the well-trained CNN in direct n umerical simulation (DNS). Substantial drag reduction rates up to 19 % and 16 % are obtained based on the spanwise and streamwise wall sh ear stresses, respectively. Furthermore, we explore the online control of wall t urbulence by combining the RNL model with reinforcement learning (RL). The RL is constructed to determine the optimal wall blowing and suction based on its obse rvation of the wall shear stresses without using the label data on the detection plane for training. The controlling and training processes are conducted synchr onously in a RNL flow field. The control strategy discovered by RL has similar d rag reduction rates with those obtained previously by the established method. Al so, the training cost decreases by over thirty times at Re-tau = 950 compared wi th the DNS-RL model. The present results provide a perspective that combining th e RNL model with machine learning control for drag reduction in wall turbulence can be effective and computationally economical."
BeijingPeople's Republic of ChinaAsiaCyborgsEmerging TechnologiesMachine LearningTechnologyTsinghua Univer sity
NETL
NSTL
